FIELD OF THE INVENTION
The present invention relates to an image processing apparatus, and more specifically to an image processing apparatus wherein image signals obtained from a continuous-tone image and unsharp signals produced from the image signals are mixed together using each of weighted parameters to be selected according to the level of density of the continuous-tone image, thereby to enable noise to be appropriately reduced over a wide range from a shadow area to a highlight when the continuous-tone image is reproduced using the mixed image signals.
Where the continuous-tone image is converted into an electric signal by an solid-state image sensing device such as a CCD, followed by reproduction of the converted signal on a photographic film or the like, noise may enter a reproduced image owing to operation of the solid-state image sensing device. There has thus been known an image processing method for averaging image signals as electric signals with a view toward reducing the visible level of such noise.
When level regions of total density of the continuous-tone image are averaged so as to produce the averaged result, the reduction in the visible level of noise is achieved. However, such a process develops inconvenience such as reduction in resolution at the highlight, i.e., the lightest spot or area of the continuous-tone image. In other words, the human vision has a characteristic with low spatial resolution at the shadow area of the continuous-tone image, while it has a characteristic with high resolution at the highlight. Accordingly, the noise does not appear appreciably at the shadow area, while the visible level of the noise and the resolution are reduced at the highlight.
In an apparatus which has been disclosed in Japanese Patent Laid-Open No. 63-156475, it has thus been practiced to process the above-described image signals for average only on the side of the shadow area at the density level regions of a continuous-tone image. In other words, in such a prior art, each of the image signals obtained based on the continuous-tone image is compared with a predetermined threshold value established between the shadow area and highlight and image signals at the shadow area are processed for average based on the result of comparison. At this time, images with no appreciable noise are obtained, on the side of the shadow area of a reproduced image, from the averaged image signals. To the contrary, images free from reduction in resolution are obtained because the image signals at the highlight are not processed for average.
However, in this prior art, a judgment is made as to whether or not the image signals should be processed for average by comparing each of the image signals with one threshold value or several threshold values. As a consequence, the resultant reproduced image will develop an areas with an abrupt change in smoothness or sharpness of the image, in areas at the instance of changing from the shadow area to the highlight. In particular, an image having a gentle change in the density gives rise to inconvenience that it develops unnaturalness on the image before and after the threshold value or one of the several threshold values referred to above.